Search

Deep learning (DL) is one of the fastest-growing topics in materials data science, with rapidly emerging applications spanning atomistic, image-based, spectral, and textual data modalities. DL allows analysis of unstructured data and automated identification of features. The recent development of large materials databases has fueled the application of DL methods in atomistic prediction in particular. In contrast, advances in image and spectral data have largely leveraged synthetic data enabled by high-quality forward models as well as by generative unsupervised DL methods. In this article, we present a high-level overview of deep learning methods followed by a detailed discussion of recent developments of deep learning in atomistic simulation, materials imaging, spectral analysis, and natural language processing. For each modality we discuss applications involving both theoretical and experimental data, typical modeling approaches with their strengths and limitations, and relevant publicly available software and datasets. We conclude the review with a discussion of recent cross-cutting work related to uncertainty quantification in this field and a brief perspective on limitations, challenges, and potential growth areas for DL methods in materials science.

A pit lake arises as a consequence of anthropogenic activities in opencast mining areas. These water bodies may be enriched in hazardous stable contaminants and/or in naturally occurring radionuclides depending on the local geological conditions. Mining legacy in Sweden produced hundreds of these pit lakes and most of them are used for recreational purposes in the southern part of the country. In this paper, one pit lake was selected for having enhanced levels of natural radionuclides. Physico-chemical parameters (temperature, pH, oxidation-reduction potential, dissolved oxygen and depth), elemental composition (via Inductive Coupled Plasma Mass Spectrometry) and radiometric characterization (via alpha spectrometry of 226Ra, 210Po and 210Pb) were carried along the depth of a 60 m depth pit lake, with the main aim to describe how natural radionuclides and elements behaves with depth in a non-uraniferous pit lake. Based on observed changes in physico-chemical parameters, a thermocline and a chemocline region were identified at around 10 and 30 m depth respectively. Concerning radionuclides, 226Ra ranged from 75 ± 3 up to 360 ± 12 mBq/kg while 210Po ranged from 11 ± 1 up to 71 ± 3 mBq/kg. 210Pb distribution with depth was also determined via secular equilibrium with 210Po after 2 years and also stable Pb was measured. Disequilibrium 226Ra-210Pb was found and the residence time of 210Pb in the water column was assessed. Additionally, different vertical distributions between 210Pb and Pb were found which points out different sources for different lead isotopes in the water body.

Natural radioactivity in the environment is a feld gaining more attention in last decades. This work is focused on the study of natural radioactivity complemented with elementary characterization at former non-uraniferous mining areas in Sweden. This aim is addressed through the study of mining lakes, called pit lakes, which are water bodies generated after opencast mining. Environmental matrices (water, sediments and rocks) from 32 Swedish pit lakes, commonly used for recreational purposes were radiometrically characterized via alpha (238U, 234U, 232Th, 230Th, 210Po isotopes) and gamma spectrometry (238U and 232Th series radionuclides). Additionally, ambient dose rate equivalent in the immediate surrounding of each pit lake was quantifed. Physico-chemical parameters (pH, specifc conductivity, dissolved oxygen, oxidation–reduction potential) and elemental composition (major and trace elements by ICP-MS) were analysed in water samples and elementary composition of sediments/rocks was measured by XRF and SEM–EDX in some specifc cases. A non-negligible number of pit lakes (26%) with enhanced U levels in water was found. At some sites, rocks contained up to 4% of U in areas with high degree of interaction with local population. Concerning the elementary perspective, another popular site (due to its turquoise water) was found to have elevated dissolved heavy metal levels. Results obtained in this work prove that measurement of natural radioactivity is another component that should be included in routine analysis of characterization in mining areas, especially if restauration of post-mining sites is intended for human recreational.